The Mac Revisited and Updated

7% paper presenh a new model correlation technique based on an iteration of the MAC meficient. The validation of II complex model such IIS rl complete rotating machine cm be very di#icult. It depends upon cmwct loadiwtion of modelling errors which few existing cow&ion methods hmv succeeded. ‘Ibe MAC value has been used in model correlation, but it is not suficiient to locate errors. A subsequent step to investigate the correlntion is to use the COMAC m enhanced COMAC. For 1? model which is not well correlated the results may suffer frum dificulties in mode pairing. A newly developed ranking method based on the MAC iteration procedure and presented here shows a mmv IDbust behrmiour than the MAC and COMAC. A major rerrsonfm this is that II normalisntion and nunif pairing is no longer m-cc-sspry. Another advantage of this new tool is that several weightingfinctions can be built in. All these methods can also be used Jbr modifying dyrwmic chmacteristics of a structure, and a combinntion of all these n&hods provide a toolbox which can be usedfor judging nltemutive modifications. A case study using a mmplete automotive alternator OS a complex model for the exptimpntal nnd Wle analytical model is presented in this paper Nomenclature MAC : modal assurance criterion MSF : modal scalefactor DOF : degree of freedom COMAC : co-ordinate assurance criterion EMA MSF COM MCC DV HPV N [@I